Improving the Accuracy of BCR-ABL Transcript Quantitation: Pitfalls of Using a Common Standard for b3a2 and b2a2 Transcripts

Abstract 4423

Considerable variation is observed in BCR-ABL analysis by qPCR making it difficult to compare inter-laboratory results in monitoring therapy of CML. Many factors contribute to the variation (primers, probes, and standards). The combination of two primers plus a probe and b3a2 plasmid standard from the European Against Cancer (EAC) program is widely accepted to quantitate two major types of p210, b3a2 and b2a2.

Since the amplicon size and sequences are different between b3a2 and b2a2, we used EAC primers plus probe with b3a2 and b2a2 standards to quantitate b3a2 and b2a2 respectively. It was found that with well determined copy numbers (CN) of both types of standards, the Ct (threshold cycle) values of b2a2 standards are smaller than b3a2 standards at each concentration point especially at lower CN (30000 CN, 0.26 cycle earlier; 300 CN, 0.42 cycle earlier; 30 CN, 0.56 cycle earlier compared to b3a2 standards from 32 experiments). In order to find out if these differences in Ct have any effect in quantitation of b2a2 samples, RNA samples pre-determined as b2a2 p210 were amplified and CN quantitated in the presence of b3a2 and b2a2 standards in the same qPCR plates. For b2a2 samples, a higher ratio of b2a2 to ABL using b3a2 standards was observed compared to the ratio using b2a2 standards (1.48 fold, p=0.0025, n=45). The data was further broken down into three International Scale (IS) ranges. For >1–10% IS range, the ratio by b3a2 standard is 1.30±0.08 fold (n=12) compared to b2a2 standard. Similarly, for > 0.1–1% IS range, the ratio is 1.46±0.08 fold (n=12) and for ≤ 0.1% IS range, the ratio is 1.60±0.22 fold (n=21).

Equal numbers of b3a2 and b2a2 RNA samples were analyzed by reference lab and GHSU lab to establish IS conversion factor (CF) for GHSU lab. There was a small difference between b3a2 and b2a2 CF when their own respective standards were used (0.67 for b3a2, 0.57 for b2a2). Therefore, a common CF of 0.62 can be generated for both types of p210. With application of common CF (0.62) to convert b3a2 and b2a2 raw ratios to IS, only 7% is under-estimated for b3a2 and 9% is over-estimated for b2a2. However, when b2a2 and b3a2 quantitation uses only b3a2 standard, there is a big difference in CF value (0.67 for b3a2 vs. 0.39 for b2a2), and a common CF (0.51) can be still generated when two types of BCR-ABL samples are mixed equally. The accuracy of results in IS were jeopardized if the labs used only b3a2 standard for both b3a2 and b2a2 quantitation. For example if this CF (0.51) is used to convert b3a2 and b2a2 raw ratios, 24% is under-estimated for b3a2 and 31% is over-estimated for b2a2.

These variations may look small but may be exaggerated if the labs have big variations in their assay and do not have equal numbers of b3a2 and b2a2 to establish the CF. Further, data from recent survey results were compared to see if the labs using only b3a2 standards for both types of p210 have higher ratios for b2a2 BCR-ABL than the labs using two individual standards. The results of commercially available standards (b3a2 only) were compared with GHSU results, which uses both types of standards (b3a2 and b2a2) to evaluate CN's of respective fusion gene.

In one survey with b3a2 RNA, the means of original raw ratio and ratio in IS by using the commercial reagent were very similar to the results by GHSU using b3a2 results (1.13 - raw ratio; 1.34 - IS between two). In two other surveys where b2a2 RNA was used, the means of raw ratio in samples analyzed using commercial reagent (b3a2 Standards) were 1.46 and 1.48 fold, compared to means of same samples analyzed by GHSU using b2a2 standards. Similarly, 1.59 and 1.76 fold difference is observed between samples using commercial standards (b3a2) and GHSU standards (b2a2) in the second survey. Moreover the ratios in IS reported in the second survey by commercial reagent were 1.85 fold and 2.23 fold of GHSU results. These findings suggest that there is a statistically significant over-estimation of b2a2 to control gene ratio, especially at MMR range, when b3a2 standards are used for b2a2 analysis. This may be one of the important factors that contribute to the considerable variation in BCR-ABL analysis. The phenomenon may be explained by the sequences or structure of an extra exon in b3a2 that is missing in the b2a2 fusion gene, and this may result in the different amplification rates of two types of p210. It is suggested that in order to quantitate the ratio of b3a2 and b2a2 accurately, their own standards should be used in qPCR analysis.